blew up a UCD400. Why did it die?

[mp006ltk]

I bought two complete hypex amp kits for ucd400's with hg power supplies, and their transformers, with softstart. I built both of them, and one cooked itself. The other is running great. They are both wired exactly the same, and one ran much hotter then the other, right before it died. Does anyone on here know what the most common occurances of failure are? Both power supply fuses blew, and now when I test the amp, I'm getting a short between the + and - power input. I'm worried that the softstart and power supply may also have been affected. How is the best way to make sure the PS is ok before I get another UCD 400, also any idea what casued this??? Thanks!

[BWRX]

Most common reason for failure: user induced Heck, I managed to fry my UCD180 some months back and haven't had the time to repair it yet!

Since both fuses blew you obviously drew a lot of current from the supply, which would indicate that something shorted - probably at the speaker output. Unhook the supply, replace the fuses, and load the + and - supply rails with some equal value resistors. They don't have to be high power resistors, just draw a small amount of current from each rail. If you have 45V rails a 10k 1/4W resistor loading each rail would work. Then power the supply modules back up and make sure each rail voltage is the correct value. Chances are the supply module is fine.

With your damaged UCD400 disconnected, check the resistance between the terminals of the output FETs. You'll probably see some short circuits which means the output FETs are toast. There may also be other damaged parts, as was the case with my UCD180.

[classd4sure]

There's really only a single common cause of failure since these modules are so robust..... as BWRX stated, human failure.

Overcurrent protection would have shut them down in the case of a short at the speaker, so I doubt that's the cause, but you never know.

Low impedance between rails as you're measuring on it is sure indicator of a blown output stage, so you know it's toasted! Most likely you lost a few parts of the mosfet drivers as well, since when mosfets blow it's always with all three leads shorted together. Sometimes you get lucky and the drivers survive though, 50/50 chance of that it seems.

The only way I can see this happening is if you made a mistake and reversed the power rails of the module, that will lead to a blown output stage real fast, and is about the only thing they can't take, aside from maybe submersing in water, solder bridges, having your pliers sitting across the PCB, or something obvious like that.

One trap that can happen is if you use your own Fast On connectors to accomadate a thicker wire gauge and crimp them at the wrong angle, and don't do a very good job with the heatshrink, it's a good way to cause a short, but that would be directly across the power rails and blow the fuses instantly, most likely leaving the output stage intact.

I'd take a real close look at how you wired it again, try and find what went wrong. It may not be repairable if the rails were reversed. If it turns out all was wired well and you can't see what went wrong, Hypex will probably want to have a look at it themselves.
Test the supply like BWRX said, it's probably fine.

[mp006ltk]

I have all my wireing color coded and the faston connectors are very well insulated and crimped. Both amps ran fine to start, and when I powered them down, one just felt hot. When I tred to power them bot back on a day later, only one started. The other amp was dead. I double and tripple checked my wireing, and still nothing. I could understand if I hooked something up wrong and killed it right away, but the dead unit was workign fine for about 20min. Thanks!

[tinitus]

Contact Hypex, I think they should deliver a new module

When you have tested your supply I hope you have "drained" it with resistor before connecting output module ??

[classd4sure]

Well, it can happen I suppose. Email Hypex see what they say, they may like to have a look at it. They're pretty good with this kind of thing.

[SSassen]

Well, I know all modules leave Hypex fully tested, the UcD400's are even set for lowest distortion, so in 99.9% of the cases we can rule out a DOA.

I've seen people knock off a SMD component (or two) and wonder why their modules didn't work before. I've also heard of people connecting the + and - of the power supply to the + and - of the loudspeaker output, which obviously also renders the module defective.

Best regards,

Sander Sassen
http://www.hardwareanalysis.com

[ptwining]

I had the experience SSassen relates in his post. That's how I roasted my first UCD's - miswiring.

Sounds unlikely for you though, mine blew up spectacularly almost immediately. Absolutely my fault, even though I checked everything sixty-six times and still... POP!

I still haven't gotten around to removing the remains of the blown decoupling cap and testing it tho. Too busy enjoying the new ones!

Good luck,
Tom

[classd4sure]

Well, while it's nice to give people the benefit of the doubt, I would think this module blew instantly too.

There's very little grey area when it comes to power switching between the rails at ~400kHz, it either works or it's smoke and you find out which pretty quick.

"Both amps ran fine to start, and when I powered them down, one just felt hot. When I tred to power them bot back on a day later, only one started. The other amp was dead. I double and tripple checked my wireing, and still nothing. I could understand if I hooked something up wrong and killed it right away, but the dead unit was workign fine for about 20min."

This seems to suggest it wasn't running fine, or at all, and had already blown taking the rail fuses with it by the time you turned it off.

Yet you say it ran fine for twenty minutes??? I'm not sure I'd tell Hypex this story when you're asking for a replacement module Maybe you tweaked the pot a little too much and when it shut down thought "I'll check it in a day after it cools"..

Did you touch the back of the module at any time while in operation? Can't do that either, that's probably one of the more common causes of failure I've seen and it still falls to user boo boo.

What you're trying to tell us isn't making alot of sense though.

[ptwining]

I color coded my wiring too. All + was red all - was black. All the same, power AND output. I'll never do that again!!

Now, all power IN is red and black - all power OUT is green and white. Too damn easy (for a noob like me) to screw up.

If you start messing with wiring to fix a problem late in the day - tomorrow can easily be a disaster

Hope that wasn't what screwed you up...

Not that I know anything at all, but what does your grounding scheme and back-plane look like? How did you mount the modules? It's easy to short s**t out on the chassis.

Regards,
Tom